Table 14 Credit scoring performance of the three models.
Datasets | Metrics | ECSTabNet + SRCGMDH | TabNet + ECSGMDH | ECS-SDE |
|---|---|---|---|---|
GMSC | Save | 0.44036(2) | 0.25329(3) | 0.45448(1) |
AUC-PR | 0.24319(2) | 0.20808(3) | 0.24449(1) | |
AUC-ROC | 0.80079(2) | 0.63848(3) | 0.80102(1) | |
BS+ | 0.19840(1) | 0.27091(3) | 0.21223(2) | |
BS– | 0.15295(2) | 0.15857(3) | 0.14019(1) | |
PAKDD | Save | 0.30444(2) | 0.05202(3) | 0.30556(1) |
AUC-PR | 0.23855(3) | 0.27665(1) | 0.25155(2) | |
AUC-ROC | 0.59240(3) | 0.59459(2) | 0.60951(1) | |
BS+ | 0.18705(2) | 0.42895(3) | 0.18299(1) | |
BS– | 0.41916(3) | 0.15244(1) | 0.38199(2) | |
DCCC | Save | 0.33076(2) | 0.30172(3) | 0.33307(1) |
AUC-PR | 0.43992(1) | 0.41095(3) | 0.41299(2) | |
AUC-ROC | 0.71824(2) | 0.69484(3) | 0.72550(1) | |
BS+ | 0.22586(2) | 0.24352(3) | 0.22276(1) | |
BS– | 0.18231(3) | 0.17478(2) | 0.17378(1) | |
IEEE | Save | 0.50054(2) | 0.47631(3) | 0.51258(1) |
AUC-PR | 0.47742(3) | 0.56474(1) | 0.50040(2) | |
AUC-ROC | 0.85593(2) | 0.81716(3) | 0.86714(1) | |
BS+ | 0.37235(2) | 0.37332(3) | 0.36915(1) | |
BS– | 0.01440(3) | 0.01409(2) | 0.01314(1) | |
Average ranking | 2.20 | 2.55 | 1.25 | |
